Data Displays That Drive Decisions

Financial reports, sales reports and operational summaries are essential. They establish what occurred and help the business meet routine control and compliance obligations. The limitation appears when a report raises a question but cannot answer it.

A monthly sales figure may show that revenue increased. It may not show whether the increase came from high-margin products, discounted stock, a wholesale customer with extended terms or a temporary promotion. An inventory report may show units on hand, but not whether those units are likely to be needed next week once weather, trading pattern, event activity, wholesale orders and recent web traffic are considered.

The distinction is practical. Reporting records a result. A management display should help the team decide what to inspect, what to change and what evidence supports the decision.

Alastair McLeod, CEO of Wallace Corporation, sees the same pattern across business systems. "The issue is rarely that a business has no information. It is that the information arrives from different systems at different times, in a form that does not answer the decision in front of management."

An interactive display cannot resolve inconsistent records. If sales appear in a POS, e-commerce platform and wholesale system, while costs, refunds and supplier payments sit elsewhere, the work begins with reconciliation. The product, customer, order and time period must mean the same thing wherever they are used.

That is why Wallace Corporation commonly approaches visualisation as the visible layer of a wider data-integration project. The first task is to identify the systems that hold the relevant information, establish ownership when values disagree, and create a dependable analytical dataset. The display then becomes a way of investigating that dataset.

A hypothetical hospitality operator illustrates the point. The management team has POS transactions, weather data, wastage records, wholesale orders, web traffic and social activity. None of those sources alone can explain tomorrow's demand. Consolidated, they can support a model that estimates likely volume, identifies the drivers of that estimate and lets management test how a different forecast would affect purchasing, staffing and production.

The value is not that a model predicts perfectly. The value is that management can compare the prediction against its current planning method, see the evidence that influenced it and make the final call with more context.

A margin explorer begins with revenue and shows the successive components that produce a final contribution result: cost of goods, payment fees, shipping, returns, discounts and other relevant operating costs. Management can select a period, product category, channel or customer type to see where the change occurred.

The interaction matters because the first view should raise a question. A fall in margin may be caused by freight, refunds, supplier pricing or a product mix change. Selecting the affected component exposes the underlying orders and records rather than producing another aggregate chart.

Demand planning simulator

A demand-planning display combines the signals that management already uses informally. For a hospitality, retail or wholesale operation, those may include historical sales, seasonality, weather, events, pre-orders, wastage and digital demand signals. It shows a forecast alongside the recent baseline and allows management to test a controlled scenario such as a promotion, an event or a supplier constraint.

This does not turn planning into autopilot. It gives the person responsible for stock, staffing or production an explicit estimate, the factors behind it and a way to compare the proposed action with the cost of being wrong.

Exception and customer-value view

A customer or account view can show patterns that are difficult to see in a monthly report: a valuable account whose purchasing has slowed, an account with high revenue but costly service interactions, or a customer segment that responds well to a particular offer. Filters let the team inspect the relevant period, channel, product group or account history.

The purpose is not surveillance or a generic score. It is to direct attention to a manageable set of cases where a relationship, pricing decision or service response may need review.

The design question is operational

The most common mistake in visualisation work is choosing the chart before choosing the decision. A map is useful when geography changes the answer. A line chart is useful when the time sequence matters. A scatter plot is useful when management needs to compare margin, volume and velocity across many products. A complex display is not automatically better than a short exception list.

Tyrell puts the operational test simply: "If a manager sees the display and still has to ask someone for the spreadsheet, it has not shortened the decision. The display should make the next question clearer and show where the answer came from."

What a business should expect

A good implementation should define the management question, data sources, refresh cycle, ownership, access rules and review process before visual design begins. It should also make the limitations clear. A forecast is an estimate. A customer score is an input to judgement. A margin figure is only as reliable as its cost allocation and reconciliation process.

The businesses that benefit most are not necessarily those with the largest datasets. They are the organisations where management repeatedly spends time assembling information, debating which number is right or reacting after a change has already become expensive.

The next action is to establish the data systems before commissioning the display. Identify the systems of record, reconcile their definitions, assign ownership and build a clean analytical layer that management can trust. Without that layer, visualisation is decoration around disputed numbers. With it, a display can expose the decision, the exception and the evidence needed to act.